New preprint on how temperature and humidity affect the infectiousness of SARS-CoV-2 and other viruses.

Key finding: the virus remains infectious longer at low temperatures and extreme relative humidities.…
What do I mean by "extreme" relative humidities (RH)? There's a U-shaped pattern: the virus does well at low RH and high RH, and less well at intermediate RH
We didn't just want to report this; we wanted to understand it. We took mechanistic principles identified by @linseymarr, @Lakdawala_Lab, and others (…) and encoded them into a simple mathematical model that we could fit to our data
Details of the model are in the paper, but the key *idea* is the following: virus inactivation follows the principles of chemical reactions
Reactions proceed faster in hotter solutions (since more collisions between particles get over the "energy barrier") and in more concentrated solutions (since collisions are more frequent, all else equal).
Ambient temperature sets the temperature of the solution in which a virion finds itself—no surprise there!
But ambient relative humidity sets the concentration. Why? Water evaporates from the solution until a (quasi-)equilibrium with the ambient environment is reached. Lower RH --> more evaporation...
...but below a certain threshold relative humidity, the efflorescence relative humidity (ERH), an electrolyte solution like cell culture medium or human respiratory fluid can crystalize. Concentrated solutions are favorable for reactions. Crystals may be less so
So we expect that U-shaped pattern. Above the ERH, more humidity --> better virus survival. Below the ERH, virus survival should rise again. And that's exactly what we see!
A simple (4 main parameters) mathematical model encoding these ideas fits our data extraordinarily well. Again, this is not a regression; it's our mechanistic model!
And using that mechanistic model, we can extrapolate to unobserved conditions and predict observations from other human coronaviruses—with success!
Lots more details in the paper, but some key epidemiological and basic science takeaways:
Epi takeaways: outdoors, you've got sunlight and ventilation to help you, even if it's cold.

In typical indoor settings, climate control keeps things cool and dry. Add in poor ventilation and limited UV, and virus does pretty well
And in specialized indoor settings like a food processing plant, it should do even better…
Further underscores what we already knew: workers need better PPE, better ventilation, testing, and paid sick leave
Basic science takeaway: simple chemistry can explain a lot about virus environmental stability, and parameters from SARS-CoV-2 predict other human coronaviruses well.

Conclusion: shared simple mechanisms may affect virus environmental stability for a number of enveloped viruses
That's it for now. Thanks very much to anyone who made it all the way through the thread, and hope you'll check out the paper!

P.S. As always, our code and data are open for anyone who wants to play around
Addendum: speaking of open code and data, we used a number of wonderful free and open source software tools, including @mcmc_stan, #Rstats, the #tidyverse (esp #ggplot2), and @mjskay's wonderful tidybayes and ggdist packages. Grateful to the communities that maintain them!
Another footnote, since people are wondering about humidities: the ERH of 45% that we use in the figures is a best guess for NaCl dominated electrolyte solutions. But practical ERHs can vary a bit. So don't assume 45% = efflorescence and 46% = solution

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Dylan Morris

Dylan Morris Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @dylanhmorris

6 Oct
There are several empirically successful models for reducing the medical, social, and economic costs of COVID until the vaccine
1) crush it and get back to normal (Vietnam, New Zealand, Taiwan)
2) Mask up + test, trace, and isolate (South Korea)
3) Mask up + cluster bust (Japan)
There are also a few models that have had high-profile failures:
1) Let it burn while shielding the vulnerable (failed in Sweden)
2) Wait until the right moment and then intervene (failed in UK)
There's also "shut down but not enough to crush it, then rapidly reopen", but I'm not sure I'd dignify that by calling it a model
Read 5 tweets
26 Sep
This pandemic should upend how biologists think about virus transmission—for the better. My greatest regret is having parroted received wisdom early on. I said our @NEJM aerosol stability findings were important for HCW 1/n
But I added that there was "currently no evidence" that everyday people in everyday settings needed to worry about aerosol spread. I publicly changed my tune after the Skagit choir cluster, which I saw—and still see—as strong evidence 2/n
In my work, I make mechanistic mathematical models of biological systems. Experiments are an important part of science, but so are theory, mechanisms, and models. We can often draw strong inferences from known physics, chemistry, and biology before we even do a new experiment 3/n
Read 22 tweets
24 Sep
I'm outraged. My doctor friend was sent into a poorly-ventilated patient room where an unmasked man screamed at her. She was advised not to wear her N95, and wore a surgical mask. He has tested positive. She's waiting for her result, but has sore throat and loss of taste. 1/2
But go on, please tell me again that discussions of transmission route are merely semantic. Healthcare workers need engineering controls, PPE, and masking protocols for patients that reflect our current knowledge
One more point: I think here there is also an error of dichotomization at play. The emerging evidence that surgical masks—and even cloth masks—can help as PPE gets translated into the false impression that they're just as good PPE as an N95.…
Read 4 tweets
17 Aug
Whenever I criticize COVID's-about-to-be-over grifters, I feel the need to say that not all optimists are grifters. Far from it. Nor do I myself thing everything is doom and gloom. I see many reasons for cautious hope:
1. faster testing
2. improved ventilation / engineering controls
3. increased PPE supply
4. last but very much not least, heartening progress toward a working vaccine
The reason I am critical of the grifters is that they propose evidence-free and/or illogical reasons why we shouldn't press forward with (1) through (4) above. Hope is important. But it should inspire (re)commitment to making that hope a reality through judicious action
Read 4 tweets
12 Aug
I should be working, but this seems to be necessary. A quick thread on R_0, R_e, the difference between them, why both are (useful) abstractions, and why neither is a fixed quantity in time or place

*deep breath*
Infectious diseases can spread because, while infectious one person with the disease can infect others. For most diseases people eventually recover (or die) and cease to be infectious. In the interim, some infectious individuals will infect more people, others will infect fewer
But will the disease spread in our population of individuals? To know that, we need to know what's going to happen in the aggregate. On average, how many new infections does each infectious individual cause before recovering or dying?
Read 19 tweets
31 Jul
The tone of articles like this one is so frustrating.

Vietnam's approach *has* been effective. Many, many lives have been saved. And no, it's not going to "catch up" in its death toll before the vaccine… Image
The media needs to stop grading on a curve. Another example: "maybe Sweden's approach worked after all" stories due to Sweden's weekly average death rate settling in at ~12x that of Denmark Image
(Per capita. Even worse in absolute terms; Sweden is about 2x larger than Denmark)
Read 4 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!